Everyone probably agrees that individuals vary in "brightness" or "intelligence," in their capacity for adaptive thinking and action. Beyond this agreement, however, there exists considerable controversy about the extent to which these variations in intelligence are genetically determined, the extent to which they remain constant throughout the life cycle, and the point in development at which such variations become measurable and predictable entities. In recent years articles about genetic differences in IQ and about IQ "fixity" have appeared frequently in the popular news media as well as in the scientific journals. The related issue, of when variations in intelligence can be reliably measured, has been all but ignored in the contemporary rush to develop programs for the early "intellectual stimulation" of infants. It is to the question of the reliability and validity of infant intelligence tests that Michael Lewis and Harry McGurk address themselves in a recent article in Science